ETFs focused on artificial intelligence are popular right now, but there’s a wide gulf between ETFs that use AI to invest and ETFs that invest in companies that benefit from or work on AI technologies. Let’s drill deeper into the difference between the two.
First and foremost, it’s important to recognize what artificial intelligence is, and what it is not. For example, we may be impressed and even unnerved by the widespread use of OpenAI’s ChatGPT and all its output. But ChatGPT represents just one branch of AI and computer science: machine learning (ML). ML uses data and algorithms to imitate the way that humans learn, iterating on its process over time and through human interaction to become more accurate.
Despite the breathless headlines from otherwise reputable news sources, however, ChatGPT lacks the self-awareness and the ability to reason that a “true AI” may one day possess — but that doesn’t mean that investors can’t be excited about other branches of artificial intelligence. Far from it — ChatGPT and its many cousins, from automated harvesters and robot painter’s assistants to Microsoft’s (MSFT) latest version of Bing, have the potential to drive markets forward.
That’s where AI ETFs come in, investing in areas of the economy that can benefit from enhancements from across the AI family tree. For example, investors can invest in the ARK Autonomous Technology & Robotics ETF (ARKQ), which has returned 19.6% YTD, investing in firms poised to benefit from AI advancements, ranging from Deere & Company (DE) and its AI-powered harvesters to Tesla (TSLA) and its work on self-driving systems.
The actively managed Franklin Exponential Data ETF (XDAT) offers a different look, investing in global firms in the big data space that may benefit from advancements in AI processes like machine learning. XDAT has returned 7.9% YTD itself, charging 50 basis points and holding companies like Cloudflare (NET), which offers security for web apps routed through its global network, which could explode in the wake of ChatGPT’s big debut.
Finally, there’s the Defiance Quantum ETF (QTUM), which tracks the BlueStar Machine Learning and Quantum Computing Index and invests in firms involved in the research and development of quantum computing and machine learning. Quantum computing refers to hardware and software designed to use superfast computers which rely on quantum physics, a frontier of AI technology with a great deal of potential. QTUM has returned 12.1% YTD, charging 40 basis points.
ETFs Powered by AI
On the other hand, the investing community has access to ETFs powered by AI. These do not offer exposure to the AI theme, but instead offer exposure to a strategy that uses AI as a method for stock selection.
“AI-powered funds take the human element out of stock picking and provide a more active approach than what many advisors expect out of smart beta rules-based strategies. Artificial intelligence-based ETFs can move faster and learn more while humans sleep,” Todd Rosenbluth, head of research at VettaFi, said. “But these funds often fly under the radar.”
ETFs powered by AI have not garnered significant assets to date and are expensive compared to plain vanilla ETFs. One of the cheapest offerings charges 58 basis points, with the majority of the funds in the space charging over 100 basis points.
“A lot of the biggest earliest use cases in AI and machine learning have been in the trading and hedge fund spaces, whether it’s using natural language summaries of filings or trolling Twitter for sentiment,” Dave Nadig, VettaFi’s financial futurist, said. “I don’t actually believe there are a lot of secret patterns in markets that can be exploited for more than very short periods of time.”
Launched in October 2018, the AI Powered Equity ETF (AIEQ) was the first actively managed ETF to fully utilize AI as a method for stock selection. Other broad equity ETFs powered by AI include the BTD Capital Fund (DIP), the AdvisorShares Let Bob AI Powered Momentum ETF (LETB), and the WisdomTree International AI Enhanced Value Fund (AIVI).
Two other AI-powered ETFs, the Merlyn.AI Bull-Rider Bear-Fighter ETF (WIZ) and the Merlyn AI SectorSurfer Momentum ETF (DUDE), use fund-of-funds structures. WIZ and DUDE each use an AI algorithm to analyze momentum indicators and assess market risk, adjusting exposures accordingly.
The Teucrium AiLA Long-Short Agriculture Strategy ETF (OAIA) is a quantitative long-short strategy used to seek to achieve market-neutral exposure to the global agriculture market. A market-neutral strategy seeks to profit from both increasing and decreasing prices; OAIA’s underlying index uses AI to make allocation decisions, generally comprising between one and nine agricultural commodities futures contracts.
Nadig said that most of the modern AI-as-portfolio managers systems become a form of black-box smart beta, where they start with a universe and use a set of rules that are machine learned to narrow that universe. These funds become interesting, however, when they show a real performance edge — something Nadig said he hasn’t seen demonstrated yet.
“That doesn’t mean AI doesn’t have a role in the investment process, but much like ChatGPT or Stable Diffusion, the real use cases are as human augmentation, not human replacement,” Nadig added.
For more news, information, and analysis, visit the Disruptive Technology Channel.